What „VR Therapy” Leaves Out
That question exposes the real limitation of thinking about VR, or eye tracking, or AI, as separate point tools rather than as parts of a connected system. Each one, evaluated on its own, tends to solve a narrow problem well, while leaving the rest of the clinical workflow untouched.
Take documentation as the clearest example of this pattern, because it is the function every clinician already recognizes as a daily source of strain. A 2025 scoping review in the Journal of Evaluation in Clinical Practice found that poorly designed electronic health record interfaces disrupt clinical workflow through task switching, prolonged screen navigation, and information fragmented across the record, and that clinicians frequently resort to workarounds such as duplicating documentation just to get through a visit (Olakotan et al.). That review also cites a striking data point: each one-point drop in a system’s usability score has been associated with a three percent increase in burnout risk (Olakotan et al.). Rehabilitation therapists are not exempt from this. A 2024 qualitative study published in the Journal of the American Medical Informatics Association found that outpatient physical, occupational, and speech therapists experience documentation burden similar to what has already been documented among physicians and nurses, with manual data entry taking time away from the patient, extending work into after-hours, and contributing to burnout (Schwartz-Dillard et al.). None of that changes because a clinic also owns a VR headset. If the exercise data a patient generates in a session still has to be typed into a separate note by hand, the technology has added a new source of engagement without removing the administrative load that was already wearing clinicians down.
The same fragmentation shows up on the assessment side. Eye tracking has produced genuinely useful clinical findings. A 2024 study in Cerebrovascular Diseases found significant correlations between eye tracking metrics and cognitive test performance in post-stroke patients, with clear differences in saccade velocity and gaze path velocity between those with and without cognitive impairment (Chan et al.). A 2026 systematic review in JMIR Rehabilitation and Assistive Technologies concluded that eye tracking shows real potential as an objective, low-burden approach for quantifying cognition-relevant behavior after acquired brain injury, particularly for inhibitory control and predictive attention. That review was also careful to note that current evidence does not yet support using eye tracking to replace conventional neuropsychological assessment on its own („Eye-Tracking Technologies”). In other words, eye tracking is a genuinely promising measurement tool, but it has mostly lived inside research protocols and specialized labs rather than inside the ordinary rhythm of a therapy session, where a therapist needs that data connected to the exercise the patient just performed and the note that has to get written afterward.
Artificial intelligence has followed a parallel path. A 2025 mini narrative review and SWOT analysis in Frontiers in Digital Health documented that AI is already being used to build personalized treatment plans, support ongoing patient management, and adapt therapy sessions in real time, and that automating certain tasks can reduce human error and free up clinician time for direct patient care (Attoh-Mensah et al.). The same review was candid about a real limitation: the accuracy of AI-generated treatment plans and real-time adaptations still tops out around 70 percent in the studies reviewed, which means AI functions best as decision support that a clinician reviews, not as an autonomous replacement for clinical judgment (Attoh-Mensah et al.). That caveat matters, and it also reinforces the point: AI adds the most value when a therapist can see its output next to the patient’s actual performance data and outcome history, in one place, rather than as an isolated report from a separate system.
Zoom out from any single technology and a broader pattern appears across digital health generally. A 2025 study in PLOS Digital Health described how the rapid proliferation of narrowly focused apps and tools has created what researchers call app fatigue, the cumulative burden and disengagement that comes from managing too many disconnected systems and data streams, and argued that the field needs to shift from proliferation toward integration (Ali and Thu). Rehabilitation technology is not immune to that dynamic. A clinic that adds a VR system, a separate eye tracking tool, a separate AI-assisted documentation add-on, and a separate outcomes platform has not solved the fragmentation problem. It has just moved it from the software vendor’s marketing deck into the therapist’s actual workday.
